AdvMIL: Adversarial multiple instance learning for the survival analysis on whole-slide images

Pei Liu, Luping Ji, Feng Ye, Bo Fu

Abstract: Highlights•An end-to-end weakly-supervised survival analysis framework for whole-slide images.•The framework includes a conditional GAN integrated with multiple instance learning.•Existing models of WSI survival analysis can be easily integrated into the framework.•Our model can fulfill the estimation of time-to-event distribution on gigapixel WSIs.•The framework also enables a semi-supervised application of WSI survival analysis.